Ontology-Based genes similarity calculation with TF-IDF
ICICA'12 Proceedings of the Third international conference on Information Computing and Applications
| Hi-index | 0.00 |
Although semantic similarity has been broadly applied in artificial intelligence and related fields, the calculation of such similarity still remains a great challenge, appealing for the development of effective methods that can be flexibly applied to a diversity of domains. In this paper, we first review existing methods that rely on an ontology to calculate semantic similarity. We classify these methods into three categories: methods based on the structure of an ontology, methods based on the information content of an ontology, and methods that utilize multiple properties of an ontology in a hybrid manner, and we analyze the advantages and limitations of these methods. Then, we propose a novel method called DOPCA that relies on the structure of an ontology to calculate semantic similarity. Our method combines two similarity measures, the degrees of overlap in paths (DOP) and the depth of the lowest common ancestor node (DLCA), and uses their weighted summation to quantify the relatedness of terms in an ontology. We apply our method to the gene ontology (GO) and the plant ontology (PO), and we show the well agreement of our method with two existing methods. Finally, we show that our method is capable of overcoming the limitation of existing methods that overlook the existence of multiple lowest common ancestor nodes, and we analyze the flexibility of our method when applied to ontologies of different domains.